- 143

The name of the Fijian beachcomber Charles Savage is for some people a byword for the savagery and ferocity which are sometimes alleged to have characterised the first generation of European settlers in the Pacific Islands. This reputation of Savage's began to be recorded in print in the 1830s, about 20 years after his death; as time passed so did his putative savagery increase. According to the Reverend Thomas Williams, usually a reliable source for Fijian history, Savage was one of a group of degraded beings so depraved that even the cannibals regarded them as monsters (Williams 1870: 3-4); and another self-proclaimed authority called them “demons in human form” after contact with whom the “vindictive and cruel” Fijians became even more depraved. (Thurston 1881 :37).

Many later historians have echoed this view, and, despite some recent revisions, historians can still be found asserting that “Savage soon became the most notorious beachcomber in the South Seas” (Gunson 1977: 102). Any attempt to discover the truth about Charles Savage—his character and his role in Fijian history—is frustrated by the nature of the sources. Those which are contemporary with him say little; those which say most are remote from him by between 15 and over 100 years. They are mutually inconsistent, and most of them make assertions which are at variance with the usual beachcomber experience in Fiji. The “documentary sources” are, in fact, oral sources which became written down decades after Savage's death, mostly by people who had not known him personally. These sources are not, therefore, direct evidence of Savage's life and times any more than oral evidence usually is.

Some scholars have recently been moved to publish warnings about the incautious use of oral sources collected by historians (King 1978, Oliver 1978, Ward 1980).1 These warnings point out that oral sources which have to be renewed with each generation are subject to two principal sources of error to a greater degree than written sources: the uncontrolled imperfections of the human memory, whether of participants, witnesses or later recipients and transmitters of information; and the deliberate modification of information - 144 through suppression, embellishment, or falsification for personal or political advantage, or for some other cultural reason. Written sources can be subject to the same kinds of distortion, and are not in principle inherently more reliable. All evidence of the past, written or oral, must be subjected to the same processes of rigorous analysis and evaluation. This warning is apposite to oral sources which have become fixed in print, of the kind that exist for Charles Savage. The purpose of this paper is to survey the use which has been made of the Savage evidence, to analyse that evidence, and to venture some suggestions about what can be known with certainty about this man.

The several stories of Charles Savage are built around a common core: he arrived in Fiji in 1808 as a survivor of the wreck of the American brig, Eliza. In due course he came under the patronage of Naulivou, the vunivalu ‘paramount chief’ of Bau, one of the most powerful states of Fiji. He acquired high status and some influence as a warrior with the aid of firearms which, according to some, he introduced into Fijian warfare. Most of the stories about Savage are concerned with this role, and the rewards it earned for him. The degree of consensus is low. It is commonly suggested that Savage, with his firearms, was responsible for Bau's eminence. Savage's career continued for five years: he died in 1813 in the “Dillon's Rock Affair” described by Peter Dillon (1829: I, 12-24). He had gone to Bau to work for the sandalwood traders, and was killed in a battle between them and the Fijians.

Savage's career in Fiji thus began in much the same way as many other beachcomber stories. After it finished it became the foundation for a legend which distinguished Savage from all other Pacific Island beachcombers in his stature as a warrior, his status as a Fijian chief and his consequence as a man of mark in Fijian history. The legend itself is merely the skeleton of a story full of romance and drama, beginning with shipwreck and lost treasure in uncharted seas, among islands whose inhabitants were known only for their ferocity, its hero a resourceful sailor who defied the cannibal ovens and the caprice of cannibal chiefs to become a chief among cannibals himself. His elevation to notable warrior and powerful chief, with a retinue of high-born wives and personal attendants, his career culminating in a death which had all the elements of high tragedy, is the substance of many an unrecorded fantasy and akin to a mass of adventure literature now no longer in vogue.

To the writers of Fijian history this story has been an allegory for the historical impact of Western civilisation on Fiji. He became in retrospect a medium of, and symbol for, both social change and political transformation. It is alleged that he fostered scepticism of Fijian mores; that his military and political activity made Bau the political centre of Fiji, and therefore the principal centre of European attention. Bau's eventual failure to govern Fiji made it possible for some to see Savage as the ultimate cause of Britain's - 145 annexation of Fiji in 1874. This role, the European as deus ex machina in Pacific history has been repeatedly challenged by modern scholarship, so that whereas in the past Savage could symbolise European potency and penetration, he is now due to become a symbol of no more than an outmoded colonial outlook.

The principal contributions to a revision of Savage's role have been made by Dorothy Shineberg and Peter France. France showed that the alleged political consequences of Savage's career were not consistent with Bau's known circumstances at the time; Bau was already before 1808 one of the pre-eminent Fijian states (France 1969: 21). Shineberg argued that claims made for Savage's impact were unrealistic inasmuch as they depended on reports of marksmanship which were not compatible with the technical characteristics of firearms of the early 19th century (Shineberg 1971: 78-9). Both these revisions conflict with the earlier work of H. E. Maude, who, although cautious of his sources, was not prepared to dismiss their claims altogether. He concurred in the judgment of the more temperate of them that Savage played an important contributory role in the rise of Bau (Maude 1968: 158-9). France, Shineberg and Maude, however, did not use all the known sources in their assessment of Savage.

That this cautionary work of academic historians has not been noticed by popularising historians is evident in the account by Brown in his history of Fiji, Men From Under the Sky (1973).2 Brown is typical of the non-academic historians in that he regarded his sources as authorities. Thus, conflict between his sources was not taken as an indication of unreliability but as complementarity. Where there was an open contradiction between the two sources on which he drew, he simply presented the one which seemed to him to be most plausible without warning the reader that there might be a problem. Brown's principal sources were Thomson, South Sea Yarns, (1894) and Tatawaqa, “Charlie Savage” (1912-1913), but he also drew on Im Thum, The Journal of William Lockerby (1925), Seemann, Viti: An Account of a Government Mission (1862) and Dillon, Narrative and Successful Result of a Voyage . . . (1829). Two of Brown's “facts” are untraced, and were apparently his own invention. He concluded that Savage's role was unique among beachcombers, and that while his role was significant, the context of Bauan affairs was equally so (Brown 1973: 71-87).

This interpretation bears the imprint of decolonisation, and it is this fact which presents the principal contrast with the interpretation of R. A. Derrick whose History of Fiji (1946) is still the major reference work. According to Derrick, “Savage and his kind changed Fiji, ushering in a new era: a period of muskets and ball, of civil wars, rebellions, invasions, massacres.. . .” Savage was a conquering but depraved hero. This interpretation was little more than a sequence of ancedotes and judgments which can be traced to five - 146 sources: Cargill, Memoirs of Mrs. Margaret Cargill (1841), Erskine, Journal of a Cruise . . . (1853), Waterhouse, The King and People of Fiji (1866), Tatawaqa (1912-1913) and Thomson (1894). Like Brown, Derrick regarded all his sources as equally plausible and did not state the criteria on which he selected some items for inclusion in his account (Derrick 1946: 44, 45, 55).

Most of the other modern accounts of Savage in Fiji are gossipy versions in travel books. To discuss them all would serve little purpose. Three, however, purport to be historically reliable: those of Sir Alan Burns, Fiji (1963), E. S. Dodge, New England and the South Seas (1965), and Olle Strandberg, Tigerland and South Sea (1953). Dodge's version was evidently written without reference to sources and is so much at variance with all other accounts that it may be dismissed without further discussion (Dodge 1965: 101-2). Burns' account, based on the most readily available printed sources, used the most-agreed-upon anecdotes and concluded,

Burns' Savage was not a romantic figure; it was the musket which was really important, and someone had to be the first bearer and user of it. Savage's personal influence on Fiji and its inhabitants was in Burns' view, “wholly evil” (Burns 1963: 51-54). Strandberg's book comes into the genre of travel literature, but with a difference, for he purported to be presenting new information on Charles Savage, some of it from oral sources, and some from documentary sources. Strandberg wrote that when visiting some of the smaller Fijian islands in about 1950 he found that people still sang traditional songs about Charles Savage:

The second discovery reported by Strandberg was the alleged identification of Savage as a ship's carpenter named Kalle Svensson, of Uddevalla in Sweden. Strandberg claimed to have met in Suva a Captain Kerrigan who had a copy of the Eliza's muster roll which identified the ship's carpenter as Charles Swenson, from “Udewala”, Sweden (Strandberg 1953: 224-6). Brown, for one, accepted the identification, and, because his book has probably sold reasonably well, future versions of the Savage legends can be expected to follow Brown's lead. The identification is, however, erroneous, even - 147 if Strandberg did see the document he claimed, and if the Eliza (not an uncommon ship's name) was the same one that was wrecked near Nairai in 1808. According to Patterson, one of the Eliza's crew, Savage used the name “Charles Savage” when he joined the vessel in Tonga, and had there been a vacancy for ship's carpenter and Savage filled it, one might suppose that to be the sort of fact which Patterson would remark upon (Patterson 1817: 80). Moreover, Savage—under that name—has been traced back to Sydney in the pages of the Sydney Gazette between 1805 and 1807, and Delano recorded that a Charles Savage (probably the same man), deserted his sealing gang in 1804 (Walter 1974: 67-9). Scepticism of Strandberg's cannibal songs is also due; if not invented by Strandberg, they are probably of recent origin as betrayed by the use of the full name “Charlie Savage”. Beachcombers were usually known either by their surname, or given name, or a nickname, or a name given by their hosts. That Savage was not known by his full name “Charlie Savage” is shown by Cary's report of a conversation with Savage's patron, Naulivou, in 1828 or 1829.

Evidently neither Naulivou nor Savage's successors knew him by any other name but Charlie.

Most writers whose main concern is for a good story rather than strict historical veracity add something to the tradition which they receive; Dodge, Brown and Strandberg all make some addition, and so does an account published in Pacific Islands Monthly in 1937. This last is conventionally sensational, adding that Savage boasted of being the father of 150 children. An unidentified, and untraced “old writer” is cited as the authority for this statement (Anon. 1937:5).

Apart from Walter, no writer of the last 50 years has contributed new information about Charles Savage upon which any reliance can be placed. The difference between the more recent academic historians and the others is their approach to their sources, not in the basic data. They all rely on data written or published before about 1920, and which for present purposes can be called primary in the sense that they draw directly on oral accounts. These sources can be grouped in various “traditions”. Within and between the traditions there are cross-influences as later writers drew on whatever earlier written accounts were available. First among the various traditions are those which can be called “first hand”, because they were written by people who knew Savage. There are only two; Patterson (1817) and Dillon (1829). Next chronologically are those which derive from the white settlers at Levuka. Most of these accounts came directly from David Whippy, the earliest and - 148 most prominent of the Levuka beachcombers. These accounts, which did not come immediately from Whippy, probably stemmed from him ultimately. This group includes the accounts by Cary (Wrecked on the Feejees, 1928), Eagleston (bêche de mer journals, 1834), Wilkes (Narrative of the United States Exploring Expedition, 1845), Lyth (notebook, n.d.), Erskine (1853) and an anonymous narrative published in Harper's New Monthly Magazine in 1853 (and referred to hereafter as Harper's). The third group may be called the “Rewa tradition” as recorded by Cargill (1841) and Jaggar (n.d.). Fourth is the “Bau tradition” of Tatawaqa (1912-13) and Toganivalu (n.d.). Other sources are less easily classified because they drew on unidentifiable sources or because they combine accounts from differing traditions, or because idiosyncracies in their accounts set them apart from the others. Among these are Waterhouse (1866), Turpin (1870-1896), and Thomson (1894). Thomson should be classified perhaps as an independent tradition; the others can be labelled as belonging to the “Seemann tradition” or “later settler tradition”. Seemann departed from the early settler tradition, and had a strong influence himself on later writers, namely Waterhouse, De Ricci (1875), Turpin and the Cyclopedia of Fiji of 1907. The story of Brett's Guide to Fiji (Thurston 1881) bears no discernible relationship to anything else, the truth least of all, and is not discussed here.

A notable feature of this classification is the absence of a “missionary tradition”. Two of the missionary accounts, Lyth and Jaggar are unpublished, and so did not influence later versions; moreover, Jaggar's probably came directly from the Rewa chief, Cakonauto (“Phillips”), while Lyth's was copied from the Rev. John Hunt, who had it from the white settlers at Levuka. The Rev. Thomas Williams paid no attention to Charles Savage as an individual, branding all the early European residents as escaped convicts, too intent on the unrestrained enjoyment of depraved pleasures to be of any consequence (Williams 1870:3-4). Cargill's accounts have some things in common with Jaggar's, but also differ quite significantly, and Waterhouse's account has so little in common with Cargill's as to be considered as an independent account. 3 The missionaries seem not to have wasted time at their infrequent district meetings swapping beachcomber yarns.

The relationship between the different accounts may be represented diagramatically, as on the accompanying chart (Figure 1).

The obvious difficulty with all these accounts is their chronological remoteness from their subject. Patterson's book was the first to appear, in 1817. Although this date was four years after Savage's death, Patterson was unaware of his later career. Patterson's usefulness is that he shows that Savage was already an experienced beachcomber before going to Fiji and provides a firm basis for reconstructing Savage's Fijian career. He joined the Eliza in Tonga with one John Husk. Husk and Savage claimed to be survivors

- 149

FIGURE 1

Evolution of the Charles Savage Legend, 1828-1928. Underlined names represent terminal branches; these sources were not subsequently used within the period considered here. Dillon (1829), who is not included, was the universal donor; most sources betray familiarity with his book at first or second hand.

- 150

of the Port-au-Prince cut off by the Tongans in 1806. Savage definitely was not, but with this explanation of his presence in Tonga, together with the story that the chief “Torki” was planning to take the ship, both men were given a place on board. 4 It was probably Savage's knowledge of Tongan which was such a help to the castaways when the Eliza was wrecked on Mocea reef, near Nairai—Patterson says he could speak the language of the people of Nairai and was very useful as an interpreter (Patterson 1817:86). The men of the Eliza were dispersed quickly. Captain Corey with four men went immediately to Bua for help; by the time he returned Patterson had himself gone to Batiki, and Corey was able to rescue only one more of the crew, a boy. As a result of the wreck all but seven of the crew remained in Fiji: one was drowned, and six were rescued almost immediately. The residue probably amounts to about 20. Most of the dollars and all the possessions of the castaways were lost on landing (Patterson 1817:83-4).

Dillon, who might have told us a great deal about Savage, significantly tells very little. But it is Dillon who first says that Savage was Swedish, and there is no independent verification. Dillon's account leaves no doubt that Savage was well known to the Fijians, but, until Savage's death, does not differentiate him from the other whites resident at Bau. He was merely one of

The Bauans had themselves recently killed three of these men for their ill-conduct, and only the intervention of the Bauan king saved the others, who in May 1813 were being employed by Dillon's commander, Robson, working in the boats getting sandalwood. When Robson sent an expedition to punish the Fijians for what he considered a breach of faith, the Bau whites were members of the party. In the ensuing fighting they were forced to withdraw to a low hill or rocky outcrop. Savage was one of three from Bau who had survived the fighting, together with Dillon and two more of Robson's crew. In the entire account Savage remained undistinguished until he began to make imprudent suggestions for getting off the rock. The first suggestion Dillon forbade; of the second he washed his hands; Savage descended to the crowd and after a few minutes' delay during which attempts were made to entice the others to come down, Savage was killed by being thrust face downwards in a well, and immediately was dismembered for cooking (Dillon 1829: 3,7,15, 17-8).

Dillon, by good luck and presence of mind eventually extricated himself from his awkward position, and next day tried to recover the remains of the - 151 third mate, who was one of those killed. All flesh and bones had been consumed, however, except for the first mate's thigh bones, which he was told, would be made into sail needles. Subsequent writers have stated monotonously that Savage's bones were made into sail needles, apparently incorrectly following Dillon's account. It is significant that Dillon says so little specifically about Savage: he was merely one of the crowd from Bau, imprudent, fluent in Fijian, resident five years, and his death evoked no comment beyond bare description. Dillon's sparse testimony opens the possibility that Savage was less of a hero than later stories indicate.

Patterson's statements about Savage seem to have gone unnoticed by writers on Fiji until Im Thurn published extracts from his book in his edition of Lockerby's Journal more than a century later (Im Thurn 1925). Dillon, directly or indirectly, was the source for most of the established facts about Savage which appeared in most of the accounts before Im Thurn. These “established facts” are Savage's name and nationality, the date and circumstances of his arrival in Fiji, and the date and circumstances of his death. Some of these facts, like the story about the sail needles, were frequently miscopied.

With the “early settler” tradition Savage's career as a warrior hero began. The earliest version we have of this tradition is the account by Cary, quoted in part above, and Savage's elevation as a shaper of events, an all-conquering hero, is thus attributable not to Cary's unnamed European informants, but to Naulivou himself: “He told me Charlie . . . was a great warrior and conquered all the islands.. . .” (Cary 1928:30). 5 Cary's story departs from those of Patterson and Dillon in some details; he named the vessel as the brig Eliza, gave no date, and says Charlie was the only one of the crew who remained in Fiji, that he took three or four muskets to Bau where he was a great marvel as the first white man seen there. Naulivou reluctantly allowed him to join in warfare, and so the first demonstration the Bauans had of the power of firearms was when Savage shot an opposing chief in battle. This one shot turned the battle, and earned for Charlie promotion to “head chief” with “command of the whole tribe”. He conquered the entire group, but was a stern leader, and would shoot men for trivial offences. The claims for his impact are incredible, but otherwise the story is not implausible. Its simplicity is notable, but perhaps most interesting are its contrasts with Patterson and Dillon only 15 years after Savage's death. By that time the only white men who had been contemporaries of Savage and were still alive were probably living at Rewa.

It was to be 100 years before Cary's narrative was published and, therefore, it had no influence on the further evolution of the Savage legend. It also has remarkably little in common with the rest of the early settler tradition. The common elements are the name of the brig, Eliza, Savage's personal courage,

- 152

Selected Details of the Early Settler Tradition

- 153

and his having raised Bau to a position of prominence. The early settler tradition generally is silent on the matter of Savage's own elevation; only one other source (Harper's) says that he was the first white man on Bau; only one other says that he had a musket on arrival on Bau (Waterhouse, who as a missionary was stationed on Bau from 1849); and only two others agree that he punished offences. In Cary's view these included trivialities, in the other accounts (Eagleston and Waterhouse) they were the practices of widow strangling and cannibalism.

The early settler tradition is by no means monolithic. Of 22 variables in Savage's career mentioned by the eight early settler versions, they approach unanimity on only one—and that is that he raised Bau to a position of unprecendented and irreversible prominence in Fijian politics. At the same time many of the features which are generally supposed to be a part of the Savage story are entirely absent from the early settler accounts collectively. Some of these details are summarised in Table 1.

These features can best be demonstrated if the accounts are paired by date, since one might suppose that accounts nearest in date would be most likely to agree. First Cary (about 1829) and Eagleston (1834) coincide on only three matters: that Savage was a man of great personal courage, that he punished offences and that he elevated Bau. They disagree on whether Savage was the first European to come to Bau and each mentions a few details about which the other is silent. Wilkes and Lyth can be paired next, both deriving from around 1840, but they touch only in asserting that Savage and his kind taught the Fijians the use of musketry and that Savage raised Bau to prominence by his exploits. Waterhouse, Erskine and the Harper's Magazine versions should be perhaps taken as a group; the last two both date from 1849, and were published in 1853. Waterhouse's book was not published until 1866, but he was resident on Bau from 1849. The three agree only that Savage arrived on an American brig which was wrecked (correctly named by Erskine and Waterhouse), and that Savage elevated Bau. Two of them (Erskine and Waterhouse) agree on the manner of Savage's death, but differ over its date; two of them (Harper's and Waterhouse) that he had an aristocratic wife, that he was cruel and wicked (both following Wilkes here), but the same two disagree about whether he was Bau's first white man. Erskine is silent on these matters, but does say that he exacerbated existing Fijian vice and taught musketry. Contemporary with these accounts is the muted report by Captain Worth of H.M.S. Calypso that Savage did much in teaching the chiefs how to enlarge their authority, and to bring conquered districts into entire subjection (Worth 1848).

Waterhouse can be considered again with Turpin, since Turpin arrived in Fiji in the year that Waterhouse's book was published, 1866. There is no clear evidence that Turpin knew Waterhouse's book but there is a slender - 154 possibility of their accounts being approximately contemporary. They agree about the Eliza, that Savage became leader of a retinue of whites (but disagree on number); against Waterhouse's view that Savage suppressed cannibalism, Turpin's evidence suggests that he could not. (Turpin 1870-1896: 95). On all other points they are at variance. Taking the eight accounts collectively, the level of agreement is low and appears to be random. The differences between them suggest that on the most generous assessment there was a highly individual degree of variation on what each writer selected to record; it is probable that all or most of them only ever heard snatches of anecdotes associated with the name Charlie, and it is possible that there never was a connected “Savage legend”. The variation would also be consistent with the hypothesis that a composite character called “Charlie” was invoked whenever a hero for an anecdote was needed.

The early settler accounts are, however, linked in another way—the unanimity of silence on some of the details which might be thought basic to the Savage legend. The silence covers Savage's being personally awarded honours (except Cary), the manner in which he travelled to Bau (except Waterhouse) and became established there, the possibility of his being a convict (except Lyth), the fate of the victims of his campaigns, the existence and disposal of his children, the initial demonstration of the power and use of muskets, any tactical response on the part of Bau's enemies to the new weapons (except Waterhouse, who says that on one occasion mats were held up to block the musket balls), and the circumstances of his going to Bua. They are generally silent also about the fate of the other men from the Eliza. Most importantly they are generally silent about any personal attributes or characteristics which might have given Savage individuality.

The early settler accounts, therefore, closest to Savage in point of time and deriving from resident Europeans well placed to tap the memories both of older beachcombers and of Fijians, do not constitute an accepted, consistent Savage legend. Even less, then, do they amount to anything resembling a corpus of knowledge and therefore should be dismissed as sources of biography. The lack of consensus and the direct implication of some such as Wilkes and Lyth, suggests that Savage was not a towering hero; he was one among many, and is known by name probably because he was singled out for special attention by Naulivou, the Vunivalu of Bau. We have no evidence why he merited such attention, but it was not uncommon for some beachcombers to enjoy a special relationship with a chief, presumably on the basis of personal liking.

It is the “Rewa tradition”—next chronologically—which first gives specific details of Savage's career. This tradition has only two representatives, Cargill and Jaggar, and both were written down independently and directly from Fijian oral sources. They are generally compatible, and for several

- 155

- 156

reasons are probably more accurate than the various early settler accounts, despite the fact that Savage did not live at Rewa, and that they relate to events of 30 years before. (see Table 2). Jaggar's account has never been published; Cargill published his version in 1841 in a form which differs slightly from the one recorded in his diary. The two Cargill accounts match quite closely Patterson's account of Savage's arrival in Fiji. Cargill was told the story of what happened “about thirty years ago” (i.e., before December 1839), shortly after Savage's arrival, and therefore correctly inferred a date of “about 1808” for the shipwreck of an English vessel (wrong) on Mothe [sic] Reef near Nairai (correct). The survivors came ashore in four boats (wrong), and one was killed (wrong, according to Patterson [1817:83]) for refusing to give up his clothes. They came ashore with axes and coin (correct) which they buried (Patterson says cutlasses, and burial of these treasures was not possible). After one month on Nairai (probably less) they went to Bau by canoe and were welcomed by Naulivou. From Bau some went to Rewa. They assisted Bau in war; one such campaign was against Kasavu (in the Rewa delta) where Savage fired from a canoe in the river adjacent to the town, through the reed fence. The slaughter was so great that the inhabitants sheltered behind piles of corpses and the flow of blood stained the river bank. Savage was a good marksman, was well supplied with ammunition, and usually carried a rifle. He was personally ambitious, had 30 (or many) wives of whom the principal one was the chiefly Kapua. He opposed cannibalism, and occasionally shot people for practising it. He was killed in Dillon's affray at Wailea and some of his bones were made into sail needles. He had a daughter, still living in 1839 at Rewa. This is the essence of Cargill's story, based on both of his accounts. It is generally temperate and although both the scale of the slaughter at Kasavu, and the use of the term “rifle” is puzzling, it is not otherwise implausible. An interesting contrast with the European settler versions is that there is no suggestion of his being a “king” or chief among the Fijians. Jaggar's story, obtained independently, matches Cargill's closely in places, but is more confused—for instance, that Savage came on a ship called the Charlie Queal which was wrecked on Mothe [sic] at Bukatatanoa near Nairai. 6 Again we have four boats and one man killed for keeping his clothes; removal to Bau; participation in wars, in particular a campaign to Lau (which neither Cargill nor any other accounts mention); that Savage had about 30 wives of whom the principal was Kakua, daughter of Roko Tui Bau. 7 Savage was involved in a squabble at Bau (the “yam affair”, of which there are other versions including Dillon's and Lyth's) in which the other whites were killed. Savage fled to Rewa, and thereafter lived at Rewa and Bau a month about. He had a daughter, Maria, still living, and a son, dead. The account of his visit to Bua and death, from a man who was there, differs slightly from other known accounts.

- 157

The differences between Jaggar and Cargill are such as to imply that Jaggar's is the conflation of a number of different stories; the alternation between Bau and Rewa suggests that Savage's name is being applied to a Rewan beachcomber who has been overshadowed in the creation of legend. Comparing Cargill and Jaggar the similarities seem to be more important than the differences, which are not, on the whole, incompatible. Both are silent as to his nationality; they agree on his importance to Bau; both use the puzzling term “rifle” to describe his weapon; and agree that he taught the use of muskets; they agree on the basic details of his arrival and agree that he discouraged cannibalism. Jaggar's story of an expedition to Lau is perhaps the explanation of Naulivou's telling Cary that “Charlie . . . conquered all the islands” (Cary 1928:30). The presence of details, plus this hint of agreement is perhaps an indication that the versions drawn from Fijian oral tradition are much more reliable than the other orally derived versions. These accounts taken together differ substantially from those which derive from white informants in that they provide specific anecdotes, and provide some personal details about Savage as a man rather than as a label for a set of abstractions.

The development of the Savage legend as it is known to later writers takes a new form with the version of Berthold Seemann, whose book was published in 1862. His work was surprisingly influential considering its inaccuracies, and was heavily relied on by several later writers, who can for convenience be labelled the “Seemann or later settler tradition”. Seemann's own account is brief, but he brings together most of the previously published snippets—Thomas Williams, Erskine and Wilkes—and adds details and interpretations not previously known. These innovations are probably fabrications, although Seemann probably did have the opportunity to collect Fijian traditions. For these reasons it is perhaps appropriate to paraphrase him. Bau, he noted, was said to owe its supremacy to its prior use of firearms, introduced by Charles Savage, but that this fact alone was insufficient, for Bau became a naval power (Seemann 1862:79). Savage was not a convict but a castaway, possessed of redeeming qualities, and the acknowledged leader of the other white residents. He acquired influence at Bau through teaching the use of (note not simply using) muskets. It was not indolence that kept Savage and his men from rising to absolute dominance in Fiji (here countering Williams' assertions), for Savage was astute and ambitious. To serve his ends “he exacted all the honours paid to exalted chiefs” and acquired as wives the most highly born women available. Further rise in power, however, was blocked deliberately by the Fijians who, seeing the influence he could come to wield through manipulating the vasu privileges of his sons, saw to it that none should live. Savage resisted this policy strenuously but in vain. He died in 1814 (following Erskine's error) in a war on behalf of a sandalwood vessel, the Hunter, was eaten, and his bones provided sail needles which were distributed to - 158 commemorate the victory. The misuse of Dillon is obvious; the principal innovations are his personal ambition, his insisting on certain honours and privileges, his policy of acquiring aristocratic wives and the fate of his sons (Seemann 1862:406-7). Unlike the Rewa tradition, Seemann has in common with the early settler tradition the use of general remarks rather than anecdotes of particularities. His errors, his innovations and his phrasing were echoed by the later writers who constitute the “Seemann tradition”. These are Waterhouse (1866 and discussed above), De Ricci (1875), Turpin (unpublished and probably about the same time, also discussed above) and the Cyclopedia of Fiji of 1907. (See Table 3 for a comparison of Seemann, Waterhouse and Turpin.) Each of these accounts of course has echoes of other sources as well: Waterhouse tapped the early settler tradition, and had probably read Cargill, and copied a phrase and an error from Wilkes (i.e., that Savage, a “notorious rascal” arrived in 1809); De Ricci plagiarised Thomas

- 159

Williams, then quoted Wilkes, and thereafter followed Seemann slavishly. For example, where Seemann wrote,

De Ricci wrote,

This latter formulation was copied by Turpin and the Cyclopedia. The Cyclopedia is but a pastiche of Seemann, Waterhouse, Cargill and De Ricci; Waterhouse and Cargill providing details not included elsewhere. Substantially, the Savage legend as it appears in all these accounts and transmitted to later writers, is the legend as created by Seemann. Thus hearsay and invention become authenticated history.

As influential as Seemann's version was that of Basil Thomson, who must surely have known all the sources which had been published before 1890 except perhaps Patterson and Harper's. These included Dillon, Wilkes, Cargill, Erskine, Seemann, Waterhouse, De Ricci and Thurston. He implies that he attempted to remain detached from all of them when he came to write his own version, published in South Sea Yarns in 1894. He summarised existing knowledge in these words:

We can probably take Thomson's word for it that this is a fair summary of what most European residents of Fiji believed about Savage around 1890. Thomson claimed to be getting to the “real story”:

In making this explicit claim to veracity, objectivity and the direct use of Fijian tradition, Thomson sets his version apart from all others. It shares the - 160 characteristics of the Rewa tradition of being concerned with anecdotes which are highly particular, rather than the more generalised statements which characterise the European traditions. The details are, however, different from those of the Rewa tradition of Cargill and Jaggar. For his sources he refers to “the historians of Bau”, so he probably based his account on those of several Bauan informants; the likelihood that he consulted non-Bauan authorities is also high, as he had ample opportunity to do so. Briefly summarised, his account is as follows: Savage went to Bau alone and unarmed, but returned to Nairai for his armory. He demonstrated the musket but refused to go to war against Verata until the Europeans at Verata came to Bau. In the assault on Verata Savage fired through the palisade from a canoe (cf. Cargill and the Kasavu battle), and was stabbed storming the town. His musket was the only one, and was tabu to all. For contributing to the conquest of Verata he was given the title Koroi-na-Vunivalu, and this is the first time the term is used in the Savage traditions. Savage scorned the customs associated with his honours, but his irreverence was tolerated. He was given two wives, both high-born, and one the daughter of the Vunivalu. He possessed a status and influence denied to the other whites, but his children, except a daughter Maraia, were all destroyed at birth.

Nakelo was the next objective in Bau's campaigns, and this time Savage used a sinnet litter, so that he could fire into the town at close range without being exposed to enemy projectiles. Three more towns soon fell in similar fashion and others sent tribute so that Rewa, at last, stood alone. On one occasion when Savage was away with the army the other Europeans insulted their hosts in the “yam affair” and all but two were killed in the consequent reprisal (Jaggar, Lyth and Dillon give versions of this incident). These details are not of the kind found in the early settler tradition, and possibly come from Thomson's Fijian informants. Thomson's own intrusions are clear:

Therefore, he went to Bua to work for the sandalwood ships and his daughter remembered his departure; the account thereafter follows Dillon. When he died so did his achievements and the memory of him became blurred, its deficiencies made good by legend. In conclusion, Thomson attributed to Savage the lasting hegemony of Bau which therefore made him ultimately responsible for the eventual cession of Fiji to Britain.

- 161

Thomson was a yam teller. He had limited respect for the veracity of oral tradition, and, it seems, not much respect for veracity at all. He combined the stories he heard despite their occasional inconsistencies and wove them together with some embellishment, some concession to popular taste (wives, tribal chief, and his ultimate political impact) to make a thrilling story. His version is not thereby worthless, and is at least probably closer to the truth than the Seemann tradition with which it shares some features such as the fate of Savage's children. It also has a superficially authoritative touch because of his claim to be using oral tradition, his use of long, putative quotations, references to “half-forgotten epics”, and his oratorical style is meant to echo a dignified, well-versed traditionalist. For all these reasons as well as for its ready availability, Thomson's account has been frequently drawn upon by later writers, and might indeed have acquired a misplaced recognition as the “Bau tradition”, the most authoritative version of all.

There are however, two further accounts which have a better claim to being the “Bau tradition”—by Tatawaqa, (1912-1913) and Toganivalu (ca. 1927, unpublished). (See Table 2.) There are substantial points of agreement between these two accounts, and while there are certain similarities with Thomson and perhaps other accounts, the differences are significant enough to suggest that they can be safely regarded as independent. For example, if Tatawaqa was familiar with Thomson's account he would not have made the mistake of identifying Savage's ship as the Josephine, nor would his account of Savage's death be at variance with all other accounts. Other differences suggest an independent oral tradition—for instance where Thomson has Savage in a protective litter from which he could shoot through the palisades at close range, Tatawaqa has him on an elevated platform so that he could shoot from a safe distance, over the palisade into the crowded defenders.

Toganivalu might or might not have been familiar with Tatawaqa's published account; that he did know other published accounts is clear from his reference to the Eliza and Savage's Swedish nationality. On minor details they differ—for example, on the means by which Savage sent a message to the Europeans at Verata, inviting them to join him at Bau. More importantly they agree that Savage arrived on Bau without a musket; that after obtaining one (by means which contrast in detail) he demonstrated its use, after which war was waged against Nakelo (not Verata first as in Thomson's account). For his martial successes—notable but not incredible—Savage was honoured (Koroi-na-Vunivalu) and given two chiefly wives (Adi Kakua and a daughter of Tui Lomaloma). He had one daughter, Maria, but no known living descendants at the time when Toganivalu was writing. No mention is made of sons being compulsorily “still-born”. The versions of Savage's death differ; Toganivalu having made inquiries at Wailea satisfied himself that Dillon's account was correct, implying that Dillon's story had been challenged. Hence Tata- - 162 waqa's version—that Savage, with a party, was rushed while cutting sandalwood when armed with nothing but their tools—is probably traditional, and a plausible reconstruction since no Fijian survivors of the Dillon's rock affair returned to Bau. Both of them credit Savage with giving Bau a lasting politico-military advantage.

It is apparent from this analysis that the degree of consensus among the ossified oral sources is low if they are taken as a whole. If the various accounts are grouped as a set of traditions, however, the overall picture is less simple. The greatest degree of dissension is within that set of sources called the early settler tradition. On the grounds of this dissension it is probably justifiable to dismiss them all. Furthermore, since the early settlers did not constitute a cohesive community, and since they probably had no tradition of transmitting historical information orally and accurately, there seems no reason to select any one of these stories as inherently preferable to any of the others. For the present purpose the Seemann (or later settler) tradition may be regarded as a subclass of the early settler tradition. There is no clear means by which Seemann's account may be distilled to separate fact from fiction, and therefore the safest conclusion is to reject his version as well.

There remain the Rewa tradition, Thomson, and the Bau tradition. The Rewa tradition is the closest to the events chronologically, and being fixed early in the contact period is very likely to be free of “feed-back” influences from the early white residents or visitors. Moreover, since the two accounts, collected independently, share some common errors and common facts, they may represent a reasonably reliable Fijian tradition. Thomson's account and the Bau tradition are more difficult to evaluate because of their remoteness from their subject, and the higher possibility of “feed-back” influences. Thomson's almost certainly contains more truth than either the early settler accounts or Seemann, but it is also likely that his interest in a good story led him to embellish it. The Bau tradition is the most intriguing of all. The identifiable errors in these two versions suggest a minimum of influence from the published stories, while their substantial agreement on details implies a fairly consistently retold tradition with occasional inputs from published accounts.

What remains of the Savage legend, therefore, are those parts of it least well known to recent historians whether sceptical or credulous. The Rewa-Bau traditions supply details of Savage's life of which we can perhaps be fairly sure, but are not sufficiently substantial to provide the basis for a biography. He was probably Swedish, he was a sailor, working on ships out of Port Jackson and was left by one of them in Tonga, probably in 1807. The Eliza carried him from Tonga to Fiji in 1808, was wrecked near Nairai where his knowledge of Tongan proved useful, and he was taken to Bau. He discharged his obligations to his patron, Naulivou, by joining in the latter's wars with his - 163 musket, which he used to good effect. He was favoured by Naulivou, honoured, and given wives (whose identity is known) appropriate to his status. He was intolerant of some Fijian practices and like many beachcombers, might have been outspoken enough to provoke some scepticism or dissent within Bauan society. But his longevity—five years—suggests that he probably knew when to restrain both his tongue and his hand. Naulivou found him useful as a warrior and probably welcome as a companion, and, because he was the most prominent of the early European residents and died at a time after which few successors were to appear for over a decade, he was remembered in Fijian tradition.

On the basis of the relatively high degree of consistency between the fragmentary Fijian traditions which have been recorded, it can be inferred that a reasonably high degree of accuracy was preserved not only for the 30 years up to the time of Cargill and Jaggar, but also for the next 90 years to the time of Toganivalu. Thus, the accuracy of the memories of individual Fijian raconteurs and the reliability of the transmission of the story between generations can be tentatively confirmed.

Twentieth century European writers have not used these more reliable sources, but have drawn almost exclusively on the more extravagant stories such as those supplied by Wilkes and Seemann. In the absence of more scholarly histories of Fiji, the interpretations of Burns, Dodge, Derrick and Brown are those sought and used by casual readers, students, and professional historians. It is with exaggerated and fanciful accounts that the critical observations of Shineberg and France have been concerned. Much of the extravagance can be dismissed simply on the grounds of inconsistency with the general beachcomber experience: scores of wives, sitting on the “councils of the tribe”, interfering in local politics, and becoming a cannibal chief are all preposterous.

Savage's notoriety was therefore posthumous, and his prominence in myths of European authorship can now be explained. By going to Bau Savage had a head start in fame. Among contemporary Europeans there, he was perhaps the natural leader, perhaps the best fighter, perhaps the best linguist, and certainly the one most favoured by Naulivou, who was arguably the most powerful man in Fiji at the time. His relative longevity also gave him an advantage in becoming better known. Nevertheless, even with these advantages Savage nearly shared the obscurity of his fellows, as Cary's scanty remarks indicate. His fame began with Dillon's account of his extraordinary death. This report was published in 1829, early enough to be the basis of the inquiries of all visitors except Cary, and so it seems likely that the Savage legend was brought into being by such inquiries rather than by any local excitement which Savage was still causing decades after his death. Once the practice of telling Charlie Savage stories became established, it fed on the - 164 curiosity and thirst for sensation of a stream of visitors. Indeed, it has continued to do so, and, as recent travel books indicate, is flourishing.

The history of Europeans in the Pacific contains stories of the lives of many men placed similarly to Charles Savage: they filled similar roles and obtained similar rewards, but did not attain notoriety. What distinguished Savage was that the truth became intertwined with European fantasies about “white savages”.

Thanks are due to Dr John Young and Dr David Hilliard for their helpful criticisms during the preparation of this paper.

• ANONYMOUS, 1937. “Memories of Charlie Savage.” Pacific Islands Monthly, 8 (2):5.
• ANONYMOUS, 1853. “A Cruise After and Among Cannibals.” Harper's New Monthly Magazine, 7:455-75.

• BROWN, Stanley, 1973. Men from Under the Sky. Rutland, Charles E. Tuttle.
• BURNS, Sir Alan, 1963. Fiji. London, H.M.S.O.

• CARGILL, Rev. David, MS. 1839-1840. Diary, Vol. 2. Mitchell Library, Sydney.
• —— 1841. Memoirs of Mrs. Margaret Cargill. London, John Mason.

- 165

• [CARY, William S.], 1928. Wrecked on the Feejees. Nantucket, Inquirer and Mirror Press.

• DE RICCI, J.H., 1875. Fiji: Our New Province in the South Seas. London, Edward Stanford.
• DERRICK, R.A., 1946. A History of Fiji. Suva, Government Printer.
• DILLON, Peter, 1829. Narrative and Successful Result of a Voyage in the South Seas.... 2 vols. London, Hurst, Chance.
• DODGE, E.S., 1965. New England and the South Seas. Cambridge, Harvard University Press.

• EAGLESTON, J.H., MS. 1834. Extracts from the Log of the Mermaid.
• ERSKINE, J.E., 1853. Journal of a Cruise Among the Islands of the Western Pacific. London, John Murray.

• FRANCE, Peter, 1969. The Charter of the Land. Melbourne, Oxford University Press.

• GUNSON, Niel, 1977. “The Coming of Foreigners,” in N. Rutherford, (ed.), Friendly Islands. A History of Tonga. Melbourne, Oxford University Press. pp. 90-113.

• IM THURN, Sir Everard and L.C. WHARTON, 1925. The Journal of William Lockerby. London, Hakluyt Society.

• JAGGAR, Rev. T.J., MS. 1837-1843. Journal. National Archives of Fiji.

• KING, Michael, 1978. “New Zealand Oral History.” The New Zealand Journal of History, 12:104-23.

• LYTH, Rev. R.B., MS. nd. Notebook. Mitchell Library.

• MAUDE, H.E., 1968. “Beachcombers and Castaways,” in Of Islands and Men. Melbourne, Oxford University Press.

• OLIVER, W.H., 1978. “Oral and Other History.” The New Zealand Journal of History, 12:99-103.

• PATTERSON, Samuel, 1817. Narrative of the Adventures and Sufferings of Samuel Patterson. Palmer, The Press.

• SEEMANN, Berthold, 1862. Viti: An Account of a Government Mission to the Vitian or Fijian Islands.... Cambridge, MacMillan.
• SHINEBERG, Dorothy, 1971. “Guns and Men in Melanesia.” The Journal of Pacific History, 6:61-82.
• STRANDBERG, Olle, 1953. Tigerland and South Sea. London, Michael Joseph.

• TATAWAQA, Pita. 1912-1913. “Charlie Savage.” Transactions of the Fijian Society.
• THOMSON, Basil, 1894. South Sea Yarns. London, William Blackwood and Sons.
• THURSTON, H.C. (ed.), 1881. Brett's Guide to Fiji. Auckland, H. Brett.
• TOGANIVALU, Ratu David, MS. n.d. An Island Kingdom. National Archives of Fiji.
• TURPIN, E.J., MS. 1870-1896. Extracts from the Diary and Narratives of Edwin J. Turpin. National Archives of Fiji.

• WALTER, Michael A.H.B., 1974. “A $40,000 Question....” Journal of the Polynesian Society, 83:58-83.

- 166

• WARD, Alan (in press). “Not Just ‘Land Wars’; the arrest of Te Kooti Te Turuki, 1889.”
• WATERHOUSE, Rev. J., 1866. The King and People of Fiji. London, Weslyan Conference Office.
• WILKES, Charles, 1845. Narrative of the United States Exploring Expedition. 5 vols. Philadelphia, Lea and Blanchard.
• WILLIAMS, Rev. Thomas, 1870. Fiji and the Fijians. London, Hodder and Stoughton.
• WORTH, Captain H., MS. 1848. Letter, ADM 172/2 Item 15. Public Records Office, London.